•Tech-savviness is the strongest psycho-social predictor of ride-hailing adoption.•Privacy-sensitivity the main psycho-social deterrent to pooled ride-hailing adoption.•Ride-hailing commute trips are ...typically made alone.•Pooled ride-hailing is substituting public transit use and active travel.•Ride-hailing induced trips are often made alone and serve errand/leisure purposes.
Even as ride-hailing has become ubiquitous in most urban areas, its impacts on individual travel are still unclear. This includes limited knowledge of demand characteristics (especially for pooled rides), travel modes being substituted, types of activities being accessed, as well as possible trip induction effects. The current study contributes to this knowledge gap by investigating ride-hailing experience, frequency, and trip characteristics through two multi-dimensional models estimated using data from the Dallas-Fort Worth Metropolitan Area. Ride-hailing adoption and usage are modeled as functions of unobserved lifestyle stochastic latent constructs, observed transportation-related choices, and sociodemographic variables. The results point to low residential location density and people’s privacy concerns as the main deterrents to pooled ride-hailing adoption, with non-Hispanic Whites being more privacy sensitive than individuals of other ethnicities. Further, our results suggest a need for policies that discourage the substitution of short-distance “walkable” trips by ride-hailing, and a need for low cost and well-integrated multi-modal systems to avoid substitution of transit trips by this mode.
Graph U-Nets Gao, Hongyang; Ji, Shuiwang
IEEE transactions on pattern analysis and machine intelligence,
09/2022, Volume:
44, Issue:
9
Journal Article
Peer reviewed
We consider the problem of representation learning for graph data. Given images are special cases of graphs with nodes lie on 2D lattices, graph embedding tasks have a natural correspondence with ...image pixel-wise prediction tasks such as segmentation. While encoder-decoder architectures like U-Nets have been successfully applied to image pixel-wise prediction tasks, similar methods are lacking for graph data. This is because pooling and up-sampling operations are not natural on graph data. To address these challenges, we propose novel graph pooling and unpooling operations. The gPool layer adaptively selects some nodes to form a smaller graph based on their scalar projection values. We further propose the gUnpool layer as the inverse operation of the gPool layer. Based on our proposed methods, we develop an encoder-decoder model, known as the graph U-Nets. Experimental results on node classification and graph classification tasks demonstrate that our methods achieve consistently better performance than previous models. Along this direction, we extend our methods by integrating attention mechanisms. Based on attention operators, we proposed attention-based pooling and unpooling layers, which can better capture graph topology information. The empirical results on graph classification tasks demonstrate the promising capability of our methods.
Shared ride services allow riders to share a ride to a common destination. They include ridesharing (carpooling and vanpooling); ridesplitting (a pooled version of ridesourcing/transportation network ...companies); taxi sharing; and microtransit. In recent years, growth of Internet-enabled wireless technologies, global satellite systems, and cloud computing - coupled with data sharing - are causing people to increase their use of mobile applications to share a ride. Some shared ride services, such as carpooling and vanpooling, can provide transportation, infrastructure, environmental, and social benefits. This paper reviews common shared ride service models, definitions, and summarises existing North American impact studies. Additionally, we explore the convergence of shared mobility; electrification; and automation, including the potential impacts of shared automated vehicle (SAV) systems. While SAV impacts remain uncertain, many practitioners and academic research predict higher efficiency, affordability, and lower greenhouse gas emissions. The impacts of SAVs will likely depend on the number of personally owned automated vehicles; types of sharing (concurrent or sequential); and the future modal split among public transit, shared fleets, and pooled rides. We conclude the paper with recommendations for local governments and public agencies to help in managing the transition to highly automated vehicles and encouraging higher occupancy modes.
Being a scarce resource, airport advertising spaces/billboards have a fixed capacity and the demand largely depends on the firm’s marketing efforts. Besides selling the limited capacity directly to ...its clients, it is common for an airport advertising firm to sell a portion of the capacity, at a discount price, through advertising agencies. By splitting the capacity into two channels, the airport advertising firm encounters a capacity de-pooling effect and it forms into a competition and collaboration relationship with the agencies. In this paper, we investigate the problem of how to fight against the de-pooling effect from a supply chain perspective. Specifically, we study the interactive decisions between an airport advertising firm and an agency firm. The airport advertising firm determines the maximal capacity to sell to the agency firm, and then the two firms compete on marketing efforts with the objective of maximizing their respective profits. We characterize the effect of capacity de-pooling and devise a coordination contract to fight against such effect. Numerical experiments are conducted to evaluate the potential benefit of the contract.
•Studying a stylized game-theoretical model for airport advertising space de-pooling.•Devising a contract with subsidy and reward to fight against the capacity de-pooling.•Deriving the optimal split and marketing effort decisions under different strategies.•Capacity de-pooling may induce firms to improve or lower marketing effort levels.•The devised contract can induce a win–win–win situation for the whole supply chain.
•A new hierarchical structure consisting of multiple levels of atrous convolution layers.•A novel Cascaded Hierarchical Atrous Spatial Pyramid Pooling (CHASPP) network by cascading the hierarchical ...structures is constructed.•Global features and context information is extensively collected to form a more reasonable and accurate prediction.•The segmentation method outperforms the state of the arts.
Atrous Spatial Pyramid Pooling (ASPP) is a module that can collect semantic information distributed in different scopes. However, because of the limited number of sampling ranges of ASPP, much valuable global features and contextual information cannot be sufficiently sampled, which degrades the representation ability of the segmentation network. Besides, due to the sparse distribution of the effective sampling points in the atrous convolution kernels of ASPP, large amount of local detail characteristics are easily discarded. To overcome the above two problems, a new Cascaded Hierarchical Atrous Pyramid Pooling (CHASPP) module, consisting of two cascaded components, is proposed. Each component is a hierarchical pyramid pooling structure containing two layers of atrous convolutions with the aim to densify the sampling distribution. On the foundation of such a hierarchical structure, another same structure is appended to form a cascaded module which can further enlarge the diversity of sampling ranges. Based on this cascaded module, not only rich local detail characteristics can be comprehensively presented, but also important global contextual information can be effectively exploited to improve the prediction accuracy. To demonstrate the performance of our CHASPP module, experiments on the benchmarks PASCAL VOC 2012 and Cityscape are conducted.
•We introduce a transformer-based GAN architecture with fewer parameters.•Replace the self-attention mechanism with pooling operations.•Depthwise convolution is added to provide absolute position ...information for the GAN.•Experimental results show that our proposed GAN model has fewer parameters.
Recently, the Transformers have shown great potential in computer vision tasks, such as classification detection, segmentation, and image synthesis, etc. The success of Transformers has been long attributed to the attention-based token mixer. However, the computational complexity of the attention-based token mixer module is quadratic to the number of tokens to be mixed. Therefore, the attention-based token mixer module requires more parameters and will cause a very large amount of computation. As far as image synthesis task is concerned, the attention-based token mixer module increases the computation amount of generative adversarial networks (GANs) based on Transformers. To address this problem, we propose the PFGAN method. The motivation is based on our observation that the computational complexity of pooling is linear to the sequence length, without any other learnable parameters. Based on this observation, we use pooling rather than self-attention as the token mixer. Experimental results on CelebA, CIFAR-10 and LSUN datasets demonstrate that our proposed method has fewer parameters and fewer computational complexity.
•Proposes a model to characterize the equilibrium in on-demand ride-sourcing market.•Identifies monopoly and social optimums of non-pooling and ride-pooling markets.•Monopoly optimum, social optimum ...and second-best solutions in both ride-pooling and non-pooling markets are always in a normal regime rather than the wild goose chase regime.•Monopoly optimum, social optimum and second-best solution trip fares in a ride-pooling market are lower than that in a non-pooling market under certain conditions.•A unit decrease in trip fare in a ride-pooling market attracts more passengers due to a reduced actual detour time.
With the recent rapid growth of technology-enabled mobility services, ride-sourcing platforms, such as Uber and DiDi, have launched commercial on-demand ride-pooling programs that allow drivers to serve more than one passenger request in each ride. Without requiring the prearrangement of trip schedules, these programs match on-demand passenger requests with vehicles that have vacant seats. Ride-pooling programs are expected to offer benefits for both individual passengers in the form of cost savings and for society in the form of traffic alleviation and emission reduction. In addition to some exogenous variables and environments for ride-sourcing market, such as city size and population density, three key decisions govern a platform's efficiency for ride-pooling services: trip fare, vehicle fleet size, and allowable detour time. An appropriate discounted fare attracts an adequate number of passengers for ride-pooling, and thus increases the successful pairing rate, while an appropriate allowable detour time prevents passengers from giving up ride-pooling service. This paper develops a mathematical model to elucidate the complex relationships between the variables and decisions involved in a ride-pooling market. We find that the monopoly optimum, social optimum and second-best solutions in both ride-pooling and non-pooling markets are always in a normal regime rather than the wild goose chase (WGC) regime—an inefficient equilibrium in which drivers spend substantial time on picking up passengers. Besides, in general, a unit decrease in trip fare in a ride-pooling market attracts more passengers than would a non-pooling market, because it not only directly increases passenger demand due to the negative price elasticity, but also reduces actual detour time, which in turn indirectly increases ride-pooling passenger demand. As a result, we prove that monopoly optimum, social optimum and second-best solution trip fares in a ride-pooling market are lower than that in a non-pooling market under certain conditions. These theoretical findings are further verified by a set of numerical studies.
Histopathology is a crucial diagnostic tool in cancer and involves the analysis of gigapixel slides. Multiple instance learning (MIL) promises success in digital histopathology thanks to its ability ...to handle gigapixel slides and work with weak labels. MIL is a machine learning paradigm that learns the mapping between bags of instances and bag labels. It represents a slide as a bag of patches and uses the slide’s weak label as the bag’s label. This paper introduces distribution-based pooling filters that obtain a bag-level representation by estimating marginal distributions of instance features. We formally prove that the distribution-based pooling filters are more expressive than the classical point estimate-based counterparts, like ‘max’ and ‘mean’ pooling, in terms of the amount of information captured while obtaining bag-level representations. Moreover, we empirically show that models with distribution-based pooling filters perform equal to or better than those with point estimate-based pooling filters on distinct real-world MIL tasks defined on the CAMELYON16 lymph node metastases dataset. Our model with a distribution pooling filter achieves an area under the receiver operating characteristics curve value of 0.9325 (95% confidence interval: 0.8798 - 0.9743) in the tumor vs. normal slide classification task.
•Two types of MIL pooling filters: point estimate based and distribution based.•The family of distribution based pooling filters is newly introduced.•The first systematic analysis of different MIL pooling filters is conducted.•Distribution based filters, in principle, are superior to point estimate based ones.•Models with distribution based filters perform the best in different MIL tasks.
When human visual system (HVS) looks at a scene, it extracts various features from the image about the scene to understand it. The extracted features are compared with the stored memory on the ...analogous scene to judge their similarity <xref ref-type="bibr" rid="ref1">1 . By analyzing to the similarity, HVS understands the scene presented on eyes. Based on the neurobiological basis, we propose a 2D full reference (FR) image quality assessment (IQA) method, named mean and deviation of deep and local similarity (MaD-DLS) that compares similarity between many original and distorted deep feature maps from convolutional neural networks (CNNs). MaD-DLS uses a deep learning algorithm, but since it uses the convolutional layers of a pre-trained model, it is free from training. For pooling of local quality scores within a deep similarity map, we employ two important descriptive statistics, (weighted) mean and standard deviation and name it mean and deviation (MaD) pooling. The two statistics each have the physical meaning: the weighted mean reflects effect of visual saliency on quality, whereas the standard deviation reflects effect of distortion distribution within the image on it. Experimental results show that MaD-DLS is superior or competitive to the existing methods and the MaD pooling is effective. The MATLAB source code of MaD-DLS will be available online soon.
Power Normalizations ( PN ) are useful non-linear operators which tackle feature imbalances in classification problems. We study PNs in the deep learning setup via a novel PN layer pooling feature ...maps. Our layer combines the feature vectors and their respective spatial locations in the feature maps produced by the last convolutional layer of CNN into a positive definite matrix with second-order statistics to which PN operators are applied, forming so-called Second-order Pooling ( SOP ). As the main goal of this paper is to study Power Normalizations, we investigate the role and meaning of MaxExp and Gamma, two popular PN functions. To this end, we provide probabilistic interpretations of such element-wise operators and discover surrogates with well-behaved derivatives for end-to-end training. Furthermore, we look at the spectral applicability of MaxExp and Gamma by studying Spectral Power Normalizations ( SPN ). We show that SPN on the autocorrelation/covariance matrix and the Heat Diffusion Process (HDP) on a graph Laplacian matrix are closely related, thus sharing their properties. Such a finding leads us to the culmination of our work, a fast spectral MaxExp which is a variant of HDP for covariances/autocorrelation matrices. We evaluate our ideas on fine-grained recognition, scene recognition, and material classification, as well as in few-shot learning and graph classification.